Public Acceptability Testing on Dependent Interviewing and Dependent Sampling - AWARD

For this contract, known as Public acceptability research on dependent sampling and dependent interviewing, the Contracting Authority is seeking to procure an independent, predominately qualitative, assessment of the public's views about ONS using individuals' administrative data to improve methods for survey collections. This research is part of the Administrative Data Census Project.

The aim of the Administrative Data Census Project is to develop an alternative to the traditional 10 yearly census, by reusing administrative data that is already collected by government departments. One of the key milestones for the project is to have developed and implemented a new methodology for producing population estimates by 2020.

Due to coverage errors on administrative lists held by government departments, ONS will need to run an annual Population Coverage Survey, for approximately 1% of addresses in England and Wales. This survey needs to be designed to measure 'over-coverage' on administrative records - the extent to which persons are still registered at addresses at which they have since moved out of.

An independent methodology review undertaken in 2014 recommended that ONS explore the potential use of 'dependent interviewing' to ensure that over-coverage can be accurately measured. This essentially involves drawing a sample of addresses and supplying field interviewers with a list of names of individuals that are registered at those addresses with government departments.

When conducting the Population Coverage Survey at these addresses, the interview can then check whether residents listed on the administrative records are in fact still at the address. This involves ONS taking information from individuals' registrations out into the field, and in some cases disclosing to current residents the names of people that have previously lived at the address.

ONS needs to understand public perception around administrative data being used in this way, and how people are likely to react to field interviewers that have prior information about who lives at the addresses. ONS is planning to test this type of approach as early as April 2018.

To do this we are seeking approval at the National Statistician's Ethics Committee. Before taking this to the committee we need a better understanding of the reputational risk that ONS will incur by pursuing this kind of approach with our field teams.